The Effective Teaching of Mathematics :

A Review of Research

by Reynolds D & Muijs D (1999)

School Leadership and Management
Vol. 19(3), pp. 273-288

Abstract
An outline is given of three bodies of knowledge that are in general
agreement about the characteristics of the effective teaching of
mathematics, and which are reflected in the British government’s
National Numeracy Strategy. A fourth body of knowledge related to
the development of higher order skills is outlined, together with some
further research needs.

1
Introduction
Much controversy has surrounded the issue of the standards of
mathematics achievement in the United Kingdom. Reviews of
international evidence (Reynolds and Farrell, 1996) suggest a poor
performance, with some evidence of deterioration over time, although
performance on number has been particularly poor and that on skills
such as geometry and data handling/analysis particularly good. The
recent Third International Science and Mathematics Study (TIMSS)
(Keys, Harris and Fernandes, 1996) also showed that British
performance in mathematics was poor, whilst performance in science
at ages 9 and 14 was particularly good by comparison with other
countries, making it unlikely that socio-economic factors, cultural
characteristics such as television viewing or other environmental
features of British society were responsible, given that they would have
predisposed to low mathematics and science scores.

Given the likelihood that educational factors of some kind are

implicated, a wide array of solutions have been proposed, ranging from
the completely mixed ability, whole class teaching offered in some
primary schools in Barking and Dagenham Local Education Authority,
to the National Numeracy Strategy that was generated by the British
government’s Numeracy Task Force (Department for Education and
Employment (DfEE), 1998a and b). All these approaches are covered
in more detail elsewhere (Thompson, 1999).

Whilst these very heterogeneous solutions have been proposed

historically for the British mathematics ‘problem’, more recently there
has occurred the emergence of an agreed ‘solid centre’ of practice
which is widely seen as being effective in generating high mathematics
achievement. This position is neither ‘progressive’ or ‘traditional’ (to
use the over simple labels present in political debate) but represents a
‘both/and’ philosophy of practice, rather than the mutually exclusive
‘either/or’ orientation that has existed historically. Thus, adherents to
this position would argue for classrooms having whole class interactive

2
teaching and differentiated groups and some individual work, and
would argue for calculators to be used together with both mental
calculation and mental strategies (see Reynolds, 1999 for an outline of
the new concensus and its generation).

In this paper, we outline some of the knowledge bases that have

played a part in facilitating the emergence of the new paradigm, and
which are now available to practitioners and policymakers in schools to
help develop this practice. We deal here particularly with the evidence
from three bodies of knowledge:

• the American evidence on effective maths teaching;

• the British knowledge base on effective maths teaching;

• the British ‘professional’ knowledge base from OFSTED about

effective maths teaching.

We then go further than the bodies of knowledge that have helped to

facilitate the recent concensus on effective maths teaching and look at
the multinational evidence about more advanced teaching strategies
that can be used to deliver ‘higher order’ mathematical skills. In our
conclusions, we outline some further thinking about exactly how these
kinds of teaching skills can be developed by practitioners within
schools.

American Teacher Effectiveness Research on Mathematics

Teaching
This follows a ‘process-product’ model and originated (in the late
sixties/early seventies) in large-scale classroom observation studies,
from which teacher behaviours emerged that were found to be
correlated with higher achievement in mathematics. The majority (but

3
not all) of this research was carried out amongst primary age students.
Based on this research a number of ‘active teaching’ models were
developed that were tested in a number intervention programmes, the
most well known being the Missouri Mathematics programme in the
late seventies that we outline later. These models approximate to the
whole class ‘interactive’ model of maths teaching that is currently the
focus of British national policy (DfEE, 1998a and b), and contain the
following elements:

High Opportunity to Learn

One of the factors to most consistently and most strongly affect

mathematics test scores is ‘opportunity-to-learn’, whether it is
measured as amount of the curriculum covered or the percentage of
test items taught (Brophy and Good, 1986; Hafner, 1993, Herman and
Klein, 1996). Opportunity-to-learn is clearly related to such factors as
length of the school day and year, and to the amount of hours of
mathematics taught. It is, however, also related to the quality of
classroom management, especially to what is known as time-on-task
(i.e. the amount of time children are actively engaged in learning
activities in the classroom). Opportunity to learn is also clearly related
to the use of homework, which expands available learning time.

An Academic Orientation from the Teacher

Another highly important factor that is also connected to children’s

time-on-task, is the teacher’s academic orientation. Effective maths
teachers emphasise academic instruction, and see learning as the
main classroom goal. This means that they spend most of their time on
curriculum-based learning activities, and create a task-oriented,
businesslike, but also supportive, environment. They spend time on
academic activities rather than on personal matters, group dynamics,
socialising or free time (Brophy and Good, 1986; Griffin and Barnes,
1986; Lampert, 1988; Cooney, 1994).

4
Effective Classroom Management

Effective teachers are able to organise and manage classrooms as

effective learning environments in which academic activities run
smoothly, transitions (between lesson segments) are brief, and little
time is spent getting organised or dealing with inattention (Brophy and
Good, 1986). For this to happen, good prior preparation of the
classroom and the installation of clear rules and procedures (before or
at the start of the school year) are essential. During lessons it is better
to use positive rather than negative language to cue behaviour.
Desired behaviour must be recognised and reinforced, attention must
be stimulated, for example by regularly altering teaching strategies,
and continuous monitoring during lessons is essential. All in all,
effective teachers manage to create a well-organised classroom with
minimal disruption and misbehaviour (Secada, 1992; Evertson et al,
1980; Lampert, 1988; Griffin and Barnes, 1986; Brophy and Good,
1986).

High Teacher Expectations of the Pupils

Effective maths teachers must show that teachers believe that all
children can master the curriculum (not just a percentage of children).
They emphasise the positive (e.g. if a child is not so good in one area
such as shape and space, s/he might be good in number) and these
positive expectations need to be transmitted to the children. Teachers
should also emphasise the importance of effort, clarifying the
relationship between effort and outcomes and helping pupils gain an
internal locus of control by constantly pointing out the importance of
their own work (Borich, 1996).

A High Proportion of Whole-Class Teaching

5
Moving to the nature of the classroom organisation, American research
has found that children learn more in classes where they spend time
being taught or supervised by their teacher rather than working on their
own. In such classes teachers spend most of their time presenting
information through lecture and demonstration. Teacher-led discussion
as opposed to individual work dominates. This is not to say that all
individual work is negative, individual practice is even necessary and
important, but ineffective teachers have been found to rely too much on
pupils working on their own, at the expense of lecture-demonstration
and class discussion (Evertson et al, 1980; Rosenshine and Stevens,
1986). Research has found that classrooms where more time is spent
teaching the whole class, rather than on letting individual pupils work
by themselves (e.g. with worksheets), show higher pupil achievement
gains. This is mainly because teachers in these classrooms provide
more thoughtful and thorough presentations, spend less time on
classroom management, enhance time-on-task and can make more
child contacts. Teachers giving whole-class instruction have also been
found to spend more time monitoring children’s achievement. There
were also likely to be less child disruptions, thus again increasing time-
on-task (Good, Grouws and Ebmeier, 1983; Evertson et al, 1980;
Walberg; 1986; Brophy, 1986; Good et al, 1990; Mason and Good,
1993; Borich, 1996).

In this form of classroom organisation the effective teacher carries the

content personally to the student rather than relying on curriculum
material or textbooks to do so. Information is mainly conveyed in brief
presentations, followed by recitation and application opportunities. It is
clear that in this type of instruction the teacher takes an active role,
conveying information to the students rather than just ‘facilitating’
student learning. Use of examples is important, and effective teachers
strive to make the presentation lively, incorporating an element of
performance. Presentations of new concepts are usually brief, followed
by opportunities for practice or recitation. This kind of instruction is

6
usually called ‘direct’ or ‘active’ (Brophy and Good, 1986; Lampert,
1988; Brophy, 1986; Borich, 1996; Creemers, 1994).

Achievement is maximised when the teacher not only actively presents

the material, but does so in a structured way, by beginning with an
overview and/or review of objectives. Effective teachers outline the
content to be covered and signal transitions between lesson parts.
Attention is drawn to the most important ideas, and subparts of the
lesson are summarised as it proceeds. The main ideas are reviewed
at the end of the lesson. In this way, the information is not only better
remembered by the children, but is also more easily apprehended as
an integrated whole, with recognition of the relationship between the
parts (Brophy and Good, 1986; Lampert, 1988).

In connection to this approach to teaching the whole class, the

question of whether or not setting or streaming is necessary when
teaching in this way. The answer to that question has to be a
pragmatic one, in that, while most research does not show strong
benefits from setting (e.g. Askew and Williams, 1995; Good, Mulryan
McCaslin, 1992) setting might be necessary in a context in which the
ability levels of children differs strongly.

Heavily Interactive Teaching that Involves Pupils in Classroom

Attitudes

This focus on the teacher presenting material in an active way to

students should, however, not be equated to a traditional ‘lecturing and
drill’ approach in which the students remain passive. Active teachers
ask a lot of questions (more than other teachers), and involve students
in class discussion. In this way students are kept involved in the lesson
and the teacher has the chance to monitor children’s understanding of
the concepts taught. Individual work is only assigned after the teacher
has made sure children have grasped the material sufficiently to be
ready for it. In general, effective teachers have been found to teach a

7
concept, then ask questions to test children’s understanding, and if the
material did not seem well understood, to reteach the concept, followed
by more monitoring. Teachers must provide substantive feedback to
students resulting from either pupil questions or answers to teacher
questions (Brophy and Good, 1986; Good, Grouws and Ebmeier, 1983;
Brophy, 1986; Rosenshine and Stevens, 1986; Borich, 1996).

There is no simple picture of the most functional cognitive level of

questions. It would seem that the best strategy is to use a mix of low
and higher level questions, the mix depending on lesson goals and
subject content, more cognitively complex subjects demanding more
higher level questions. The optimal post-question ‘wait time’ similarly
depends on context. When testing basic skills a short wait time is
appropriate. However, when asking more cognitively demanding
questions, or questions that are meant to elicit group exploration or
discussion, a longer wait time is appropriate. Effective teachers have
been found to ask more process questions (i.e. questions calling for
explanations by the students), though the majority of questions asked
were still product questions (i.e. calling for a single response)
(Evertson et al, 1980; Brophy and Good, 1986; Borich, 1996).
A well known example of an effective programme of maths teaching
noted earlier is the Missouri Mathematics Effectiveness Project
conducted by Good and associates in the late 1970s (Good, Grouws
and Ebmeier, 1983; Good and Grouws, 1979). On the basis of
research findings, a primary school teaching model was designed,
which teachers were trained to implement. Lessons were structured as
follows for teachers:

a) Daily Review (approx. 10 minutes)

1. Review concepts and skills associated with previous day’s
homework
2. Collect and deal with homework assignments
3. Ask several mental computation exercises

8
b) Development (approx. 20 minutes) (introducing new concepts,
developing understanding)
1. Briefly focus on prerequisite skills and concepts
2. Focus on meaning and promote student understanding by
lively explanations, demonstrations etc.
3. Assess student competence
a. Using process and product questions (active
interaction)
b. Using controlled practice
4. Repeat and elaborate on the meaning portion as necessary

c) Individual Work (approx. 15 minutes)

1. Provide uninterrupted successful practice
2. Momentum - keep the ball rolling - get everyone involved,
then sustain involvement
3. Alerting - let students know their work will be checked at
the end of each period
4. Accountability - check the student’s work

d) Homework assignment
1. Assign on a regular basis at the end of each maths class
2. Should involve about 15 minutes of work to be done at
home
3. Should include 1 or 2 review problems

Teachers were trained to follow the active teaching method as

described above throughout. Although some teachers had problems
implementing the crucial development part of the lesson, the impact on
pupils’ mathematics achievement was impressive. A replication study
and a study using a slightly revised model in secondary schools were
equally effective (Good, Grouws and Ebmeier, 1983). Programs based
on the same teacher-effectiveness principles but with modification
suited to the particular contexts, have been implemented on a smaller

9
scale in the US, such as Griffin and Barnes’ (1986) CTP project, which
also showed good results.

The British Teacher Effectiveness Research on Mathematics

Teaching

The United Kingdom knowledge base is, by contrast to the American, a

highly restricted one, although there is evidence of considerable
contemporary policy interest in ‘teaching’ (Galton, 1995) and some
promising new research avenues being explored, particularly in the
areas of teacher’s conceptual and subject knowledge in mathematics
(Askew et al, 1997) and in variation in the effects of teachers’
behaviours in lessons (Creemers and Reynolds, 1996). Interestingly,
the British results on effective teaching parallel those from America.

The notable ORACLE study, which involved a ‘process-product’

orientation similar to the American research just discussed, showed
that teachers labelled as ‘Class Enquirers’ generated the greatest
gains in the areas of mathematics and language, but that this finding
did not extend to reading. By contrast, the group of ‘Individual
Monitoring’ teachers made amongst the least progress. It is important
to note that the more successful ‘Class Enquirers’ group utilised four
times as much time in whole class interactive teaching as the
‘Individual Monitors’ (Galton and Croll, 1980; Croll, 1996).

Further analyses (Croll, 1996) correlated the academic gain made by

different classes with different patterns of class/teacher interactions,
finding a moderate positive (0.29) correlation between the percentage
of time utilised in whole class and small group interaction and

10
children’s progress, showing as Croll (1996, p.23) notes ‘.. a positive
association of progress and non-individualised interaction’. Croll
(1996) notes two dangers in any rapid translation of ORACLE findings
into recommended practice - that the ‘whole class interactive’ teachers
differed in ways other than in their class teaching techniques, and that
teachers utilising other teaching styles (the ‘Infrequent Changers’)
which did not have high levels of whole class interaction also scored
above average in gain.

The ORACLE study also looked at the children’s ‘time on task’ (or
academically engaged time) and found that whole class interaction was
positively associated with high levels of time on task, with the ‘Class
Enquirers’ having average time on task 10% higher than other
teachers. The PACE study (Pollard et al, 1994) also notes high levels
of whole class interactive teaching to be correlated with high pupil ‘task
engagement’. Further analyses by one of the ORACLE authors (Croll
and Moses, 1988) shows a high positive correlation between time in
whole class interaction and time on task. Time in group based
interaction showed no such association. It is important to emphasise in
this context of course the close link between ‘time on task’ and learning
gain, which is one of the most replicated of the America teacher
effectiveness findings that we noted earlier.

The second important British teacher effectiveness study is the Junior

School Project (JSP) of Mortimore et al (1988), based upon a four year
cohort study of pupils from 50 primary schools, which involved
collection of a considerable volume of data on children and their family
backgrounds (‘intakes’), school and classroom ‘processes’ and
‘outcomes’ in academic (reading, mathematics) and affective (e.g. self
conception, attendance, behaviour) areas. This study reported twelve
factors that were associated with effectiveness both across outcome
areas and in specific subjects such as mathematics. Significant
positive relationships were found with such factors as structured
sessions, use of higher-order questions and statements, frequent

11
questioning, restricting sessions to a single area of work, involvement
of pupils and the proportion of time utilised in communicating with the
whole class. Negative relationships were found with teachers spending
a high proportion of their time communicating with individual pupils
(Mortimore et al, 1988).

It is important at this point to note the considerable agreement between

the American research base and the British. The American literature
emphasised the gains produced by active teaching and by interactive
teaching, and the positive effect of a high proportion of teacher time on
questioning, on communication with the class and on pupil involvement
with their work that was the teaching style of the ‘class enquirers’ on
the ORACLE study as noted above. The potency of whole class
interactive instruction is shown in both the American and British
teacher effectiveness literature, and the benefits of whole class
teaching, active teaching and of interactive teaching, are all shown in
the classic Mortimore (1988) study.

It is also important to note at this point that British research shows the
deleterious effect of children working on their own, emphasising the
need to reduce unguided practice to optimise gain, as shown in the
American research too.

The British Professional Evidence on Effective Mathematics

Teaching

There is a third body of knowledge that we can look at to discern ‘good

practice’, which is that taken from the publications of OFSTED. This
knowledge base has been generated by professionals with
considerable experience of mathematics processes at classroom and
school level, and of the mathematics outcomes related to them,
although it does not necessarily possess the tight research design of
our first two bodies of knowledge above.

12
Three reports from OFSTED seem appropriate here. The first, Primary
Matters (OFSTED, 1995), outlined a number of general
teacher/teaching factors associated with positive outcomes in general.
These included factors such as good teacher subject knowledge, good
questioning skills, an emphasis on instruction, clear objectives, good
time management and good classroom organisation.

The second study was specifically related to the characteristics of the

successful teaching of basic skills and, for our purpose here, to
achievement in numeracy (OFSTED, 1996). It outlined firstly the
characteristics of classroom processes where maths standards were
low, which included:

• too much emphasis upon repetitive number work

• too much individualisation of work
• too little fluency in mental calculation.

The report went on to outline in detail the characteristics of successful

maths teaching. This:

• provides clear structure for lessons and makes good use of

time, maintaining challenge, pace and motivation;

• includes sessions of direct teaching, with the teacher

involved pro-actively and not just when pupils are stuck;

• involves regular interaction with pupils, with the teacher

using perceptive questioning, giving careful attention to
misconceptions and providing help and constructive
response;

13
• rehearses existing knowledge and skills in order to keep
them sharp and enhance them, and encourages quick recall
of as many number facts as possible;

• uses a variety of activities on a topic in order to consolidate

and extend understanding;

The third study was that undertaken by OFSTED on The Teaching of

Number in Three Local Education Authorities (1997). Some key
conclusions of this report were (Italics added):

‘In all the schools there was a mix of whole class, group
and individual work in varying proportions during the
lesson. Some teachers were much more successful than
others in the way they used these organisational
strategies to extend and direct their teaching and to
increase pupils’ number knowledge, understanding and
skills. In the best lessons there was usually a higher
proportion of time spent teaching the class together, often
at the start and sometimes at the end of the lesson, with
individual and group work closely linked to the whole-
class work. The pupils worked on tasks individually or in
groups which reinforced or extended what they had been
taught in the whole-class time and consolidated their
learning by coming together again at the end of the
lesson. The effect of this was an obvious common gain
in core knowledge and skills which made it easier for the
pupils to help one another and progress together. For
the teacher, too, this well-structured mix of whole-class,
group and individual work not only made the lesson more
manageable, it also established a climate and a common
language for talking about mathematics which benefited
more children for more of the time.

14
 These good lessons stood in sharp contrast to the poorer
work which suffered from a distinct, common
organisational weakness, notably a debilitating over-use
of individual work, and to a lesser extent, group work.
Where these weaknesses occurred there was often an
over-reliance on worksheets and published schemes. In
other words, more often than not complex arrangements
for individual work were self-defeating; they dissipated
rather than intensified the quality of the teaching and
reduced the opportunities for children to learn.

A further advantage of whole-class teaching identified by Ofsted in

their recent review of early results of the National Numeracy Project
(Ofsted, 1998) is that it is more successful in bringing pupils whose first
language is not English up to the level of their peers, thanks to the
direct teaching of mathematics terms with which they might not
otherwise be as familiar as other children.

A Fourth Area : Research on Problem Solving and Co-operative

Group Work in Mathematics

While programmes based on a direct instruction/active teaching model

have been effective in improving mathematics achievement in basic
skills, doubt has been cast on whether this approach is sufficient for
teaching higher-order mathematical problem-solving or mathematical
thinking skills, the importance of which has received increasing
emphasis in recent years (Galton, 1995). According to Peterson
(1988), while direct instruction focusing on basic skills may be a
necessary condition for being able to develop higher order thinking and
problem-solving skills, it is not sufficient. A number of additional
classroom processes may be needed to enhance higher order thinking:
a focus on meaning and understanding in mathematics, direct teaching
of higher level cognitive strategies and problem solving, and co-
operative small group work, which we deal with here.

15
The advantages of co-operative small group work to developing
problem solving skills lie partly in the ‘scaffolding’ process, whereby
pupils help each other learn in the ‘zone of proximal development’.
Giving and receiving help and explanation may develop children’s
thinking skills, as well as helping them to verbalise and structure their
thoughts (Peterson, 1988; Leikin and Zaslavsky, 1997). By co-
operating in small groups children can share their own ways of thinking
and reflect on them and on the thinking and ideas of others. This
exchange may encourage students to engage in more higher order
thinking (Becker and Selter, 1996). Pupils thus provide assistance and
support to each other. Co-operative small groups force the
accommodation of the opinions of various members, and students
must therefore search, engage in problem solving and take one
another’s perspective. In this way students can develop an enhanced
understanding of self and other, and learn that others possess both
strengths and weaknesses. This may help students who are less able
problem solvers to overcome their insecurity about problem solving
because they can see more able peers struggling over difficult
problems. The fact that a group contains more knowledge than an
individual means that problem solving strategies can be more powerful.
This may help students see the importance of co-operation. Group
members may serve as models to one another, thus enhancing
learning-to-learn skills. Students also receive practice in collaboration,
a skill they will require in real life. (Daniels, 1993; Good, McCaslin and
Reys, 1992; Mevarech and Kramarski, 1997).

A number of authors have pointed to possible problems with this

method, however. Thus, as Good, McCaslin and Reys (1992) point out,
shared student misconceptions can be reinforced by group work.
Furthermore, students might be tempted to engage in off-task social
interaction (Goods and Galbraith, 1996). Students may also receive
differential status in groups. Some may start to perceive themselves as
having little to contribute to the group, or may find that their

16
contributions are not greatly valued. This may lead them to become
passive in the group. Small group work may then advantage high
ability students over lower ability ones. Studies have found low ability
students to be less active in small groups, in part because they
understand the task less well, and in part because student talk can
also express low expectancies of certain other students (Good,
Mulryan and McCaslin, 1992; Good and Biddle, 1988; Leikin and
Zaslavsky, 1997). One of the main problems with small group work is
the fact that small groups require far more classroom management
skills from the teacher. If not well prepared, small group work can
significantly increase the time the teacher spends on direction,
transition and nonmathematical managerial activities (Good et al, 1990;
Brophy and Good, 1986). To be effective, teachers also need to have
access to a large number of problems, further exacerbating the
difficulties involved with group work. All in all, small group work
requires a lot of preparation time for the teacher (Doyle, 1986).

However, programmes that have attempted to develop problem-solving

skills through small-group work report good results, such as enhanced
conceptual understanding and higher achievement on problem-solving
tasks (e.g. Wood and Sellers, 1997; Maher, 1991; Verschaffel and De
Corte, 1993; Leikin and Zaslavsky, 1997; Townsend and Hicks, 1997;
Goods and Galbraith, 1996).
While the benefits of co-operative small group work for enhancing
students’ problem solving skills seem important, it is clear that for
group work to be effective it is insufficient to put students into groups
and let them get on with it. A number of conditions need to be met. The
most important of these, according to Stevens and Slavin (1995), who
base their findings on a large number of studies carried out by Slavin,
are group goals and individual accountability. Group goals are
essential to motivate students to work co-operatively and thus help
their groups’ mathematics learning. Individual accountability increases
the engagement of individual students and decreases the probability of

17
‘free-rider’ effects, whereby certain students choose to remain passive
while letting other group members do all the work.

According to Good, McCaslin and Reys (1992), to make small group

work effective, teachers need to design tasks that encourage the group
to work together. The use of shared manipulatives (e.g. graphs,
calculators) can be useful. The task should engage all members of the
group, focusing on processes. Each student should think about the
problem in a meaningful way, and then participate in other students’
problem solving strategies. The task should be sufficiently challenging
for it to require collective problem solving, but not so challenging that it
causes group members to give up prematurely. Work-group tasks that
introduce a concept are often more effective than tasks that review a
mathematical idea.

As with the active teaching model discussed above, a number of

models have been proposed that incorporate co-operative small group
work. Good, McCaslin and Reys (1992) propose a model which still
relies largely on whole class teaching, as the teacher spends thirty to
forty minutes directly teaching the class. Part of this time is spent on
developing new concepts and skills with a stress on meaning, by
explanation, demonstration, illustrations and discussion. After that the
work-group task is described, using explanation, investigation of the
task, generalisation and application. Five to ten minutes is then spent
on the actual work-group task, which should reinforce or extend
concepts explained during whole class teaching, while the last five to
ten minutes are spent on review. Mevarech and Kramarski’s (1997)
model leaves more time for the actual group work. The teacher starts
by introducing a new concept to the whole class, but then proceeds
immediately to group work, wherein the pupils have to answer 3 kinds
of metacognitive questions: comprehension questions, strategy
questions (select, describe and justify strategy appropriate to solving
the problem) and connection questions (describe similarities and

18
differences between the problem at hand and previously solved
problems).

Overall, it is clear that effective small group work is a structured activity

that requires a lot of teacher effort. If the subject worked on requires
higher order thinking and problem solving skills, and is used under the
conditions described above, the effort will be worth it, however, in
terms of learning gains in ‘higher order’ areas.

In no sense, though, is the development of higher order skills to be

seen as unrelated to that of basic computational and other skills, since
performance on the two sets of skills is related (Hembree, 1992). What
is required for an optimal level of achievement across a range of
mathematical skills is both whole class interactive and collaborative
group based teaching, with perhaps Key Stages One and Two more
heavily orientated to the former and later Key Stages towards the
latter.

Conclusions

Teachers and policymakers may be excused from a degree of cynicism

as to why they should believe the research and recommendations
about the importance of whole class interactive teaching and the
‘active teaching’ model we have outlined here, when for perhaps two
decades they have been encouraged to pursue other, quite radically
different, policies. Whatever the reason for these past ‘non rational’,
ineffective policies may have been, teachers and policymakers should
be clear that all the bodies of knowledge that we have converge upon
the same set of effective policies. In international surveys, it is
societies utilising high proportions of whole class interactive teaching
that have the highest test scores (Reynolds and Farrell, 1996). In
national programmes, such as the National Numeracy Project, those
based upon whole class interactive teaching have obtained impressive

19
gains (DfEE, 1999). From the within Britain and the within American
research, it is clear that the active, whole class teaching that involves
pupils is associated with higher learning gains in basic skills. When all
evidence available converges on the same practice in terms of ‘what
works’, it would be foolish to ignore it.

It is important, though, to recognise some areas where we urgently

require further research. Firstly, there is growing evidence of ‘context
specificity’ in the precise factors associated with learning gains,
originally shown in interesting research from California, where highly
effective schools in poor catchment areas pursued policies
discouraging parental involvement in the school, in contrast to the
effective schools in more advantaged catchment areas that
encouraged the practice (Hallinger and Murphy, 1986). Whilst some
factors apply across all social contexts (such as having high
expectations of what children can achieve or ‘lesson structure’), it may
be that certain factors apply only in certain environmental contexts. At
classroom level an example might be that the factor of ‘proceeding in
small steps with consolidation if necessary’ is important for all children
who are learning to read for the first time in all contexts, whilst in the
contexts inhabited by lower social class or lower attaining children it
may be necessary to ensure high learning gain through the use of
small ‘steps’ for teaching all knowledge and not just knowledge that is
new, before moving on to other approaches.

Borich (1996) gives the following summary of teacher factors that may
be necessary to achieve high achievement gains in classrooms in two
different social settings, those of low socio-economic status and
middle/high socio-economic status. Effective practices within low
socio-economic status contexts involve the teacher behaviours of:

• generating a warm and supportive affect by letting children know

help is available;

20
• getting a response, any response, before moving on to the next bit
of new material;

• presenting material in small bits, with a chance to practice before

moving on;

• showing how bits fit together before moving on;

• emphasising knowledge and applications before abstraction, putting

the concrete first;

• giving immediate help (through use of peers perhaps);

• generating strong structure and well-planned transitions;

• generating strong structure, ground-flow and well-planned

transitions;

• the use of individual differentiated material;

• the use of the experiences of pupils.

Effective practices within middle socio-economic status contexts

involve the teacher behaviours of:

• requiring extended reasoning;

• posing questions that require associations and generalisations;

• giving difficult material;

• the use of projects that require independent judgement, discovery,

problem solving and the use of original information;

21
• encouraging learners to take responsibility for their own learning;

• very rich verbalising.

We do not know as yet the extent of any ‘context specificity’ in the

precise factors associated with gains in mathematics achievement.
However, there are enough hints of the existence of this factor to make
one wary of using undifferentiated methods in highly differentiated
school contexts.

Secondly, it may be that there will be difficulties in relating the above

bodies of knowledge to practitioners, in the following ways:

• maximum take up of such knowledge is usually held to occur when

there is ownership of the process of individual/school improvement
by individuals and schools (Fullan, 1991; Hopkins et al, 1996). How
can this be maximised in the case of mathematics teaching, when
practice has been so different to that which research suggests is
optimal?

• given the long term developmental needs of the profession, how

can teachers be enabled to be active, reflexive practitioners
involved in knowledge creation about effective practices, whilst at
the same time being given defined ‘good practice’? How can the
provision of educational foundations of good practice also enable
long term development?

Given the complexity of these issues, it is essential that the new

national programmes involved in the National Literacy Strategy and the
National Numeracy Strategy, are evaluated as ‘experiments of nature’
in order to inform these issues. The likely variation between teachers
and schools in their utilisation of the programmes can furnish an

22
authentic British knowledge base concerning teacher effectiveness,
looking at universal characteristics that may be necessary to optimise
achievement in all types of catchment areas, and at context specific
characteristics that may be necessary for effectiveness within certain
school environments. Evaluation of the National Numeracy Strategy
can further test out the validity of the effectiveness factors outlined in
this paper.

23
REFERENCES

Askew, M. and William, D. (1995) Recent Research in Mathematics

Education. London: OFSTED.

Askew, M., Rhodes, V., Brown, M., William, D. and Johnson, D. (1997)
Effective Teachers of Numeracy. Report of a Study Carried Out for the
Teacher Training Agency. London: King’s College London, School of
Education.

Becker, J. P. and Selter, C. (1996) Elementary School Practices, in

Bishop, A. J., Clements, K., Keitel, C., Kilpatrick, J., Laborde, C. (eds)
International Handbook of Mathematics Education. Dordrecht : Kluwer.

Borich, G. (1996) Effective Teaching Methods (Third Edition). New

York : Macmillan.

Brophy, J. (1986) Teaching and Learning Mathematics: Where

Research Should Be Going. Journal for Research in Mathematics
Education, 17(5), 323-346.

Brophy, J. and Good, T. L. (1986) Teacher Behaviour and Student

Achievement, in M. C. Wittrock (ed), Handbook of Research on
Teaching. New York : MacMillan.

Cooney, T. J. (1994) Research and Teacher Education: In Search of

Common Ground. Journal for Research in Mathematics Education,
25(6), 608-636.
Creemers, B. P. M. (1994) The Effective School. London : Cassell.

Creemers, B. P. M. and Reynolds, D. (1996) Issues and Implications

of International Effectiveness Research. International Journal of
Education Research, 25(3), 257-266.

24
Croll, P. (1996) Teacher-pupil interaction in the classroom, in P. Croll
and N. Hastings (eds) Effective Primary Teaching. London : David
Fulton.

Croll, P. and Moses, D. (1988) Teaching methods and time on task in

junior classrooms. Educational Researcher, 30(2), 90-97.

Daniels, H. (1993) Coping with Mathematics in the National

Curriculum: Pupil Strategies and teacher Responses. Support for
Learning, 8(2), 65-69.

Department for Education and Employment (1998) Numeracy Matters

: The Preliminary Report of the Numeracy Task Force. London :
HMSO.

Department for Education and Employment (1998b) The

Implementation of the National Numeracy Strategy (Final Report of the
Numeracy Task Force). London : HMSO.

Department for Education and Employment (1999) Documents

published at detailed launch of the Numeracy Strategy, 11th January.

Doyle, W. (1986) Classroom Organisation and Management, in M. C.

Wittrock (ed), Handbook of Research on Teaching. New York:
MacMillan.

Evertson, C. M., Anderson, C. W., Anderson, L. M. and Brophy, J. E.

(1980) Relationships Between Classroom Behaviors and Student
Outcomes in Junior High Mathematics and English Classes. American
Educational Research Journal, 17(1), 43-60.

Fullan, M. (1991) The New Meaning of Educational Change. London :

Cassell.

25
Galton, M. (1995) Crisis in the Primary Classroom. London : David
Fulton Publishers.

Galton, M. and Croll, P. (1980) Pupil progress in the basic skills, in M.

Galton and B. Simon (eds) Progress and Performance in the Primary
Classroom. London : Routledge and Kegan Paul.

Good, T. L., Grouws, D. A. and Ebmeier, H. (1983) Active Mathematics

Teaching. New York: Longman.

Good, T. L. and Biddle, B. J. (1988) Research and the Improvement of

Mathematics Instruction: The Need for Observational Resources, in D.
Grouws, D. Jones (eds), Perspectives on Research on Effective
Mathematics Teaching. Reston, VA: NCTM.

Good, T. L., Grouws, D. A., DeWayne, A. M., Slavings, R. J. and

Cramer, C. (1990) An Observational Study of Small-Group
Mathematics Instruction in Elementary School. American Educational
Research Journal, 27(4), 735-782.

Good, T. L., McCaslin, M. M. and Reys, B. J. (1992) Investigating Work

Groups to Promote Problem Solving in Mathematics, in ?? (ed),
Advances in Research on Teaching, vol. 3: Planning and Learning
Tasks. Greenwich, CT : JAI Press.

Good, T. L., Mulryan, C. and McCaslin, M. (1992) Grouping for

Instruction in Mathematics: A Call for Programmatic Research on
Small-Group Processes, in D. A. Grouws (ed), Handbook of
Mathematics Teaching and Learning. New York : MacMillan.

26
Goods, M. and Galbraith, P. (1996) Do It This Way! Metacognitive
Strategies in Collaborative Mathematical Problem Solving. Educational
Studies in Mathematics, 30, 229-260.

Griffin, G. A. and Barnes, S. (1986) Using Research Findings to

Change School and Classroom Practice: Results of an Experimental
Study. America Educational Research Journal, 23(4), 572-586.

Hafner, A. L. (1993) Teaching-Method Scales and Mathematics-Class

Achievement: What Works With Different Outcomes? American
Educational Research Journal, 30(1), 71-94.

Hallinger, P. and Murphy, J. (1986) The Social Context of Effective

Schools. American Journal of Eduation, 94, 328-355.
Herman, J. L. and Klein, C. D. (1996) Evaluating Equity in Alternative
Assessment: An Illustration of Opportunity-to-Learn Issues. Journal of
Educational Research, 89(4), 246-256.

Hopkins, D. (1996) Towards a Theory for School Improvement, in J.

Gray, D. Reynolds and C. Fitz-Gibbon (eds) Merging Traditions : the
future of research on school effectiveness and school improvement.
London : Cassell.

Keys, W., Harris, S., Fernandes, C. (1996) Third International Study,

First National Report, Part One. Slough : National Foundation for
Educational Research.

Lampert, M. (1988) What Can Research on Teacher Education Tell Us

About Improving Quality in Mathematics Education? Teaching and
Teacher Education, 4(2), 157-170.

Leikin, R. and Zaslavski, O. (1997) Facilitating Student Interaction in

Mathematics in a Cooperative Learning Setting. Journal for Research
in Mathematics Education, 28(3), 331-354.

27
Maher, C. A. (1991) Is Dealing With Mathematics as a Thoughtful
Subject Compatible with Maintaining Satisfactory Test Scores? A Nine-
Year Study. Journal of Mathematical Behavior, 10(3), 225-248.

Mason, D. A. and Good, T. L. (1993) Effects of Two-Group and Whole-

Class Teaching on Regrouped Elementary Students’ Mathematics
Achievement. American Educational Research Journal, 30(2), 328-
360.
McCaslin, M., Tuck, D., Wiard, A., Brown, B., Lapage, J. and Pyle, J.
(1994) Gender Composition and Small Group Learning in 4th Grade
Mathematics. The Elementary School Journal, 94(5), 467-482.

Mevarech, Z. R. and Kramarski, B. (1997) IMPROVE: A

Multidimensional Method for Teaching Mathematics in Heterogeneous
Classrooms. American Educational Research Journal, 34(2), 365-394.

Mortimore, P., Sammons, P., Stoll, L., Lewis, D. and Ecob, R. (1988)
School Matters. Wells Somerset : Open Books.

OFSTED (1995) Teaching Quality : The Primary Debate. London :

OFSTED.

OFSTED (1996) Successful Teaching of Literacy and Numeracy in

Primary Schools : A Starting Point. For the 1996 GEST Proposals.
London : OFSTED.

OFSTED (1997) The Teaching of Number in Three Inner Urban LEA’s.

London : OFSTED.

OFSTED (1998) The National Numeracy Project, An HMI Evaluation.

London : OFSTED.

28
Peterson, P. L. (1988) Teaching for Higher Order Thinking in
Mathematics: The Challenge for the Next Decade, in D. A. Grouws, T.
J. Cooney, D. Jones (eds), Perspectives on Research on Effective
Mathematics Teaching. Reston, VA : NCTM.
Pollard, A., Broadfoot, P., Croll, P., Osborn, N. and Abbott, D. (1994)
Changing English Primary Schools? London : Cassell.

Reynolds, D. and Cuttance, P. (1992) School Effectiveness :

Research, Policy and Practice. London : Cassell.

Reynolds, D. and Farrell, S. (1996) Worlds Apart? - A Review of

International Studies of Educational Achievement Involving England.
London : HMSO for OFSTED.

Reynolds, D. (1999) Numeracy Matters : Contemporary policy issues

in the teaching of mathematics, in I. Thompson (ed) Issues in Teaching
Numeracy in Primary Schools. Buckingham : Open University Press.

Secada, W. G. (1992) Race, Ethnicity, Social Class, Language and

Achievement in Mathematics, in D. A. Grouws, (ed), Handbook of
Research on Mathematics Teaching and Learning. New York :
MacMillan.

Stevens, R. J. and Slavin, R. E. (1995) The Cooperative Elementary

School: Effects on Students’ Achievement, Attitudes and Social
Relations. American Educational Research Journal, 32(2), 321-351.

Thompson, I. (1997) The Early Years Curriculum Tomorrow, in I.

Thompson (ed), Teaching and Learning Early Number. Buckingham:
Open University Press.

Thompson, I. (1999) Issues in Teaching Numeracy in Primary

Schools. Buckingham : Open University Press.

29
Townsend, M. A. R. and Hicks, L. (1997) Classroom Goals Structures,
Social Satisfaction and the Perceived Value of Academic Tasks. British
Journal of Educational Psychology, 67, 1-12.

Verschaffel, L. and De Corte, E. (1993) A Decade of Research on

Word Problem Solving in Leuven: Theoretical, Methodological and
Practical Outcomes. Educational Psychology, 5(3), 239-256.

Walberg, H. J. (1986) Syntheses of Research on Teaching, in M. C.

Wittrock (ed), Handbook of Research on Teaching. New York:
MacMillan.

Wood, T. and Sellers, P. (1997) Deepening the Analysis: Longitudinal

Assessment of a Problem-Centred Mathematics Program. Journal for
Research in Mathematics Education, 28(2), 163-186.

30